dv6z-7000 vs. dv6t-7000

First I noticed the dv6z is much cheaper.

What are the strengths and weaknesses of each? Which one will last longer?

 

last longer? as in battery life? the dv6z.

otherwise, the dv6t is more powerful on all fronts.

you get what you pay for. dv6z is 7/10 the price for 7/10 the performance.
Actually, I'd say it's 7/10 the price for 5/10 performance, so the dv6z is actually sort of expensive

 

If you're considering spending more than $600 on a dv6z you might as well get a dv6t. Faster processor and GPU.

 

I can get 1920x1080, 7670m crossfire, A10 trinity for ~$700. Where would it lose? Would an A10 trinity be weaker than a ivy i5?

 

Don't bother with Trinity. It gets demolished by Llano and Ivy Bridge i3/i5 dual-cores in x86 performance:
AMD A-Series A10-4600M Notebook Processor - Notebookcheck.net Tech

 

sounds like you're mainly a gamer.

in that case, dv6z is fine for you, though it loses to dv6t in EVERYTHING except battery life.

dv6t destroys A10 on anything cpu related by more than double, and the GT650m destroys the 7670m, even when crossfired. GT650m is like 2x-3x faster than 7670m, and crossfire is usually an awful experience. AMD has sucky drivers.

Like i said, 7/10 the price, for half the performance.

if you don't have the money, and are ok with half performance, go for it.


I remember there was a deal for a loaded dv6t for under $1000


http://www.notebookcheck.net/Intel-Core-i7-3610QM-Notebook-Processor.72681.0.html

look at that. dv6t often has triple+ performance vs A10

http://www.notebookcheck.net/NVIDIA-GeForce-GT-650M.71887.0.html

 

There will be some gaming, but my priority will be on programming. I had no idea AMD's were lagging behind that much, that's pretty bad.

 

The best choice is A10-4600M without discrete GPU. The integrated 7660G enough for everything in 1366X near high, while CPU performance do not really matters, cannot possible feel any difference other than video encoding (if 4 core Intel). Of course Intel + Nvidia 650M combination will be faster for overall in gaming, but count with twice as much power consumption, more heat, more fan noise...

 

When you say programming, do you mean that you will compile on the laptop or just use it as a terminal to write code? If you will compile (or do anything else that is CPU-intensive), then the dv6t is the only sensible choice -- it's twice the performance for 40% more money. I have the slightly slower Sandy Bridge version of the dv6t and it is really fast -- faster than most of the Xeons and Opterons we have at work (only the recent ones are faster). I've actually shifted most of my CPU-intensive tasks from my work desktop to my laptop.

If you will only use it as a terminal, then it's not so clear cut. The display is the most important thing (don't get the 1366x768 version; it's terrible for looking at code). Beyond that, it probably doesn't matter that much. I would personally still get the dv6t, but since I am one of the few people here whose work is strongly CPU-bound, I'm biased.

 

another guy here who bought an A10 M6 said that it blows out a lot of heat too.
Reaches upper 90's under load, and triggers max fan speed.

 

That is bad, but What would happen with skinny M6 using hot head Ivy Bridge + Nvidia 650M?

 

That article is about desktops.

 

Desktops have more freedom to use better more efficient cooling systems. In a laptop, you're pretty much stuck with what the OEM made unless you do some heavy modification. In most laptops the CPU and GPU are sharing one tiny fan and a relatively small heatsink/heatpipe. Heat is still the enemy and in a laptop environment, compared to a desktop, components usually run hotter. Compound this problem with the fact that HP's laptop cooling systems haven't always been the best designs in the past or present. I understand Atom_Anti's point but I would also like to see some max temp readings for an IB I5/I7 (dual or quad) + 650M. This is one of the biggest reasons why many here on the forum are reluctant to buy the new HP Pavilion DV4 (great hardware but possible overheating waiting to happen).

 

That does not matter, all the Ivy Bridge's made with same manufacturing process works. Thinking Ivy running on higher temp than Sandy and Sandy was warmer than Llano, than Ivy won't be cooler than Trinity, even if Trinity runs warmer than Llano. +if someone wants to play games, Ivy will always need discrete GPU, which means again more heat...

 

Sandy was warmer than Llano but it also performed better. I bet if you underclocked SB so that it matched the same performance as Llano the temps would be about the same.

The same applies to IVB. An i3 should run just as cool as the A10.

 

performed better where? SuperPi? Chinebench?

 

Yeah. Do you really need to ask where SB performed better than Llano?

 

Pretty much this /
For llano, people pay 7/10 and they may get 9/10 in some area . (multithread/game blah blah blah)

I see no reason to consider trinity if you plan on getting dGPU unless it is THAT much cheaper with same performance dGPU.

Llano OC'd to i5 multithread performance pump as much heat out and still stink on many aspect.

 

Both of those, but more importantly, GCC (or any other compiler), any compression/decompression program (gzip, 7zip, WinZip/RAR -- you name it) and literally any other program which stresses the CPU.

Also, I think that you are really misrepresenting the Ivy Bridge heat issue. It is not that Ivy laptops run hotter than Sandy laptops -- they might or they might not depending on the chassis, but there is no overall trend. In fact, at stock settings, it is more power efficient than Sandy Bridge even in desktops (see the AnandTech review). The problem is that when you overclock and overvolt Ivy, you cannot push it as far as Sandy without superior cooling (note that the testing in your link was done with a substantial overclock). This has absolutely nothing to do with the behavior in laptops because you cannot overclock or overvolt normal laptop CPUs.

 

Well, I just try to understand why so important having faster CPU and how will speed up the laptop. Wide range of people are claiming Intel laptops 2-3 times faster than AMDs.......... Really ? How am I gonna feel that?
Isn't just more vise spending the extra money for SSD and good RAM?

 

you're just talking about internet browsing, microsoft office........ of course you won't feel speed.
all that stuff is not cpu intensive.

but when people need to run excel scripts, and do other hardcore cpu stuff, edit videos, etc , AMD is so slow.
Most of us with Llano wouldn't even consider buying Trinity, because we can finish multicore tasks faster than A10.

Even in gaming, AMD's "good enough" cpu philosophy is just that..... good, but not great. Extra cpu horsepower can add 20% or more frames/sec

 

I'm just pointing out that it's misleading to say IB runs hotter without taking into account that it also performs better.

If you want to talk about how you are going to feel the difference, how are you going to notice "hot head Ivy Bridge" if it spends all of its time idling while you browse the internet? The only time the processor gets hot is when it's under load. And if it's under load then CPU performance is going to matter. If you can't feel the speed difference between IVB and Trinity, then you're not going to feel a temperature difference either.

Also, you're going to feel the CPU bottleneck if you play intensive games. Like BF3. Trinity is good for budget gaming but as you move up in GPU performance you are going to be held back by the CPU.

 

Yeah, I know the CPU bottleneck is very important fact. This is what I was hearing all the time using a single core Pentium 4 3.4GHz in the mid-2000s.
"Do not upgrade with X850pro, because CPU already bottleneck the Radeon 9800pro GPU and will see no improvement, better if you change the whole computer." I did not fell to this Intel created marketing bullshet and did not give more money them by purchasing new motherboard, CPU, etc. Instead I kept the configuration and upgrade with X850pro, later X1950pro and lastly with HD3850pro. All what I could experience with these upgrades; able to play on higher resolution near higher details with newer games... (During these years could be possible change 3 times CPU and Motherboard for nothing but spending lots of money).

Meanwhile, in my comments I try to explain and give answer to bfpri question; there is no reason to choose faster CPU instead of having faster GPU and call AMD products slow. Think how often are you gonna use WinRar and H264 encode, if so; difficult to wait 10 seconds extra? Important to do faster SuperPI calculations and seeing bigger numbers in Chinebench? What difference all these makes?
I say better to see some real experiences; faster game, OpenCL acceleration, longer battery life time and money for SSD

 

With the openCL acceleration you'd actually get pretty close to QuickSync speed in encoding video. Winzip is also openCL accelerated.

For nearly everyone the difference in single-threaded performance isn't noticeable. Unless you're doing intensive CPU heavy work on a laptop (and really, you shouldn't be) then you'll find the AMD chips lack the power. For nearly everyone else, though -- and that's an overwhelming majority of people -- the boost in GPU performance and lower price tag are much more important.

Frankly, I think Trinity outpaces the IB i5's and i3's and i7's as far as laptops go. It's just a better buy in smaller laptops without discrete GPUs. If you're going to add a discrete GPU then you can jump on the Intel boat as the added CPU performance is a bonus. If you're buying anything at or below a GT540m, though, grab the Trinity A10 chip. You won't have to deal with a really hot Fermi video card.

There are definitely issues I'm having is that manufacturers are being stupid with the designs. You can't yet find a decent 14" form factor or a decent slim 15.6" one. The added GPUs are something like a 7730m which doesn't crossfire with the Trinity GPU. The laptops are also quite expensive and I have a hunch it's not because AMD is charging more for their chips but rather the makers are increasing their margins on the AMD laptops.

 

Like I already said, CPU performance matters for gaming. If you want to get Trinity without a discrete GPU, then that's fine. But you're going to be CPU limited otherwise and/or have poor performance.

I have no idea why you're ignoring gaming as a "real experience" and instead touting things like OpenCL, which is near useless for most people and goes against your whole "CPU performance doesn't matter" argument.

 

opencl doesn't use the cpu. It uses the gpu

so in this case, cpu doesn't matter...

opencl is very promising. Handbrake has a beta version that supports it. I'd like to see more software use opencl.

If Intel can't take advantage of opencl, then we have an interesting challenge.

 

OpenCL can and does run on the CPU.

But beside that, OpenCL on the GPU is used to accelerate CPU oriented programs, like Handbrake, as you mentioned. With programs like those, both the CPU and GPU matter. So faster performance can come from the CPU and the GPU.

If you care about OpenCL support, then you care about the increasing the performance of the program (i.e. the purpose of OpenCl). If you care about the performance of the program, you also care about CPU performance, because the CPU has the biggest impact on performance. At which point the whole "CPU performance doesn't matter" argument falls apart.

His (Atom_Anti's) argument was something like "who cares about waiting 10 extra seconds" to show that having a faster CPU isn't a "real experience" and then he mentions OpenCL as a "real experience" because for some reason if you use GPU acceleration those 10 seconds now matter.

Doesn't make any sense to me.

It's like the "I'm not going to notice the performance difference but IVB runs hotter" argument from last page. If you're not going to notice or don't care about the difference in CPU performance, you're not going to notice or care about OpenCL acceleration either.

 

I think his point is a little different although he certainly didn't word it correctly. Let me do that.

People focus too much on synthetic benchmarks. Though they do serve a purpose -- extrapolating as much info from the hardware such that it'll represent real world usage --, if it doesn't apply to real world scenarios then they shouldn't be thrown around as much. A perfect example is cinebench, actually. Trinity carries over the modular design and as such has only 1 FPU per 2 ALUs (AMD's "cores"). Though the FPU's perform much better than they did with Llano and support FMA3/FMA4 and AVX ISAs (and on top of that they have better throughput), cinebench is very FPU heavy and makes the architecture look much worse.

In general, heavy FPU workloads can also benefit from openCL and GPU acceleration because that's what GPU's are built for: floating point number crunching. So popular applications that are openCL accelerated and see very tangible benefits like the full Adobe line, Winzip, Sony Vegas, Maya and more which can and mostly are all be floating point heavy would show opposite results than the one in cinebench, which, remember, as a synthetic benchmark was meant to represent real world usage. It's back a$$wards.

Trinity is actually a very good chip for the midrange market. It means if you're buying a notebook with the intention of gaming you can skip on something like a GT540m or 7670m and still get the same performance within a ~10% delta (depending on the game). If you're going to add a hefty and powerful discrete GPU, though, then go with an Intel i5 + GPU.

 

Right. My point was just that doesn't make sense to say the CPU performance doesn't matter and at the same time say OpenCL perf matters. If the program being used depends on OpenCL performance, then it will also depend on CPU performance.

 

Yea. Well, obviously CPU performance matters but I'd wager that GPU performance matters just as much nowadays, particularly considering most notebooks sold are those without discrete GPUs.

As a package, if I were limited to buying a laptop with only an on-die GPU then I'd go for Trinity. I think for most people it's also the better buy. As far as CPU intensive workloads, the one they'll run into the most is probably gaming, which also happens to be the sort of scenario that Trinity trounces IB at. Most consumers use the facebooks, facetubes and what have you with an occasional game thrown in. An AMD APU does this just as well as an Intel CPU if not better and for a lower price. The performance delta between the two platforms and chips is only visible on gaming and it's going to be AMD sided. As far as high end chips and features goes, though, Intel still has a better platform. Thunderbolt (still no real Thunderbolt devices but we're getting there), SRT cache system and Quick sync are great but are also added bonuses that most people won't use -- minus SRT but only if it comes with a drive. Although I think I speak for most of us here when I say that an SSD is a better option.

As much as Bulldozer sucked on the desktop, Trinity makes a whole lot of sense Now if only we'd see some decent Trinity laptops...

 

Yeah I would totally consider replacing my current Llano/6750m setup if there was a nice A10/900p/13" setup for <~$750.

Dunno why no manufacturer has figured out how to put a 35w APU into a 13" notebook yet.

 

It's possible my English explanation not the best, therefore you did not get exactly what I mean under CPU and GPU power and power-consumption part. However mrpelo understanding me, link626 also agree with OpenCL acceleration part, so feel free reading the further comments and consider them.

 

Yea, I've been looking for much the same. I'd like to see a decent 14" or 13" Trinity laptop with an optional IPS screen, though admittedly the screen may be too much to ask for For that I may have to go with an aftermarket display.

Something like a Portege R935/945 would be perfect